Under appropriate circumstances, damage to a vehicle or other object can be estimated from a 3-D virtual model. Such estimation techniques are particularly useful where physical access to the vehicle or other object is difficult, dangerous, or time-consuming. However, the estimates of damage are limited by the quality of the virtual model, which is limited by the quality of captured data from which the virtual model is generated. Existing techniques rely upon a trained technician or other individual to perform high-quality data capture (i.e., scanning).